BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Australia/Melbourne
X-LIC-LOCATION:Australia/Melbourne
BEGIN:DAYLIGHT
TZOFFSETFROM:+1000
TZOFFSETTO:+1100
TZNAME:AEDT
DTSTART:19721003T020000
RRULE:FREQ=YEARLY;BYMONTH=4;BYDAY=1SU
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:19721003T020000
TZOFFSETFROM:+1100
TZOFFSETTO:+1000
TZNAME:AEST
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260114T163632Z
LOCATION:Darling Harbour Theatre\, Level 2 (Convention Centre)
DTSTART;TZID=Australia/Melbourne:20231212T093000
DTEND;TZID=Australia/Melbourne:20231212T124500
UID:siggraphasia_SIGGRAPH Asia 2023_sess209_papers_661@linklings.com
SUMMARY:VET: Visual Error Tomography for Point Cloud Completion and High-Q
 uality Neural Rendering
DESCRIPTION:Linus Franke, Darius Rückert, and Laura Fink (Friedrich-Alexan
 der Universität Erlangen-Nürnberg); Matthias Innmann (NavVis GmbH); and Ma
 rc Stamminger (Friedrich-Alexander Universität Erlangen-Nürnberg)\n\nIn th
 e last few years, deep neural networks opened the doors for big advances i
 n novel view synthesis. Many of these approaches are based on a (coarse) p
 roxy geometry obtained by structure from motion algorithms. Small deficien
 cies in this proxy can be fixed by neural rendering, but larger holes or m
 issing parts, as they commonly appear for thin structures or for glossy re
 gions however still lead to very distracting artifacts and temporal instab
 ility. In this paper, we present a novel neural rendering based approach t
 o detect and fix such deficiencies. As a proxy, we use a point cloud, whic
 h allows us to easily remove outlier geometry and to fill in missing geome
 try without complicated topological operations. Keys to our approach are (
 i) a differentiable, blending point-based renderer that can blend out redu
 ndant points, as well as (ii) the concept of Visual Error Tomography (VET)
 , which allows us to lift 2D error maps to identify 3D-regions lacking geo
 metry and to spawn novel points accordingly. Furthermore, (iii) by adding 
 points as nested environment maps, our approach allows us to generate high
 -quality renderings of the surroundings in the same pipeline. In our resul
 ts, we show that our approach can significantly improve the quality of a p
 oint cloud obtained by structure from motion and thus increase novel view 
 synthesis quality. In contrast to point growing techniques, the approach c
 an also fix large-scale holes and missing thin structures effectively. Ren
 dering quality outperforms state-of-the-art methods and temporal stability
  is significantly improved, while rendering is possible with real-time fra
 me rates.\n\nRegistration Category: Full Access, Enhanced Access, Trade Ex
 hibitor, Experience Hall Exhibitor\n\n
URL:https://asia.siggraph.org/2023/full-program?id=papers_661&sess=sess209
END:VEVENT
END:VCALENDAR